Add EmoNet Voice Bench results (42/42 classes, pooled r=0.304 rho=0.323) + eval artifacts

61 VoiceClap attribute-regression heads + docs, results, reference inference

initial commit